MMAudio: Taming Multimodal Joint Training for High-Quality Video-to-Audio Synthesis

TL;DR

MMAudio introduces a multimodal training framework that synthesizes high-quality, synchronized audio from video and text, outperforming existing models in quality and alignment while maintaining efficiency.

Contribution

The paper presents MMAudio, a novel joint training approach leveraging large-scale text-audio data to enhance video-to-audio synthesis and synchronization.

Findings

Achieves state-of-the-art video-to-audio quality and synchronization.

Maintains competitive text-to-audio performance.

Operates efficiently with low inference time and few parameters.

Abstract

We propose to synthesize high-quality and synchronized audio, given video and optional text conditions, using a novel multimodal joint training framework MMAudio. In contrast to single-modality training conditioned on (limited) video data only, MMAudio is jointly trained with larger-scale, readily available text-audio data to learn to generate semantically aligned high-quality audio samples. Additionally, we improve audio-visual synchrony with a conditional synchronization module that aligns video conditions with audio latents at the frame level. Trained with a flow matching objective, MMAudio achieves new video-to-audio state-of-the-art among public models in terms of audio quality, semantic alignment, and audio-visual synchronization, while having a low inference time (1.23s to generate an 8s clip) and just 157M parameters. MMAudio also achieves surprisingly competitive performance in text-to-audio generation, showing that joint training does not hinder single-modality performance. Code and demo are available at: https://hkchengrex.github.io/MMAudio